research article real-time energy management control...
Post on 08-Mar-2018
216 Views
Preview:
TRANSCRIPT
Hindawi Publishing CorporationJournal of Control Science and EngineeringVolume 2013 Article ID 801237 10 pageshttpdxdoiorg1011552013801237
Research ArticleReal-Time Energy Management Control forHybrid Electric Powertrains
Mohamed Zaher and Sabri Cetinkunt
University of Illinois at Chicago Chicago IL 60607 USA
Correspondence should be addressed to Mohamed Zaher mhzaherasmeorg
Received 19 December 2012 Accepted 9 April 2013
Academic Editor Salem Haggag
Copyright copy 2013 M Zaher and S Cetinkunt This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited
This paper focuses on embedded control of a hybrid powertrain concepts for mobile vehicle applications Optimal robust controlapproach is used to develop a real-time energy management strategy The main idea is to store the normally wasted mechanicalregenerative energy in energy storage devices for later usage The regenerative energy recovery opportunity exists in any conditionwhere the speed of motion is in the opposite direction to the applied force or torque This is the case when the vehicle is brakingdecelerating the motion is driven by gravitational force or load driven There are three main concepts for energy storing devicesin hybrid vehicles electric hydraulic and mechanical (flywheel) The real-time control challenge is to balance the system powerdemands from the engine and the hybrid storage device without depleting the energy storage device or stalling the engine in anywork cycle In the worst-case scenario only the engine is used and the hybrid system is completely disabled A rule-based controlalgorithm is developed and is tuned for different work cycles and could be linked to a gain scheduling algorithm A gain schedulingalgorithm identifies the cycle being performed by the work machine and its position via GPS and maps both of them to the gains
1 Introduction
Fuel efficiency and reduced emissions are two of the mostimportant considerations in all combustion-based powergeneration including diesel engines [1] In order to meetthese demands alternative powertrain concepts including allelectric hybrid and fuel cell are being developedThe conceptbehind the hybrid devices is to store the otherwise wastedregenerativemechanical energy in energy storage devices andreuse that energy later in the work cycle The regenerativeenergy recovery opportunity exists in every condition wherethe speed of motion is in opposite direction to the appliedforce or torque (Figure 1 [2]) This condition is satisfied invarious conditions such as (Figure 2) the following
(1) vehicle braking (ie slowdown to a lower speed onzero slope or vehicle is moving down a hill andbrakingmust be applied to maintain a desired speed)
(2) load is moved by gravitational (load) force
A parallel electric-hybrid powertrain concept is discussedin the next section The real-time control challenge is to
balance the machine power demands from both the engineand the hybrid storage device The constraints faced indeveloping the control strategy are as follows
(1) minimize fuel consumption while meeting low emis-sion requirements
(2) maintain or improve the work-machine productivity(3) prevent the depletion of the energy storage device
and maintain an acceptable state of charge (SOC)
A rule-based control strategy with multiple tunable gainsfor different work cycles is developed The control algorithmshould be robust and have low sensitivity to gain changes toaccommodate real working conditionsThis paper focuses oncontrolling hybrid powertrainmobile vehicles in real time viaan optimal robust control that will enable continuous opera-tion through repetitive and between different cycles Yafu andCheng [3] studied mild hybrid electric vehicles (HEV) withintegrated starter generator (ISG) They used a parallel assistcontrol strategy and modeled the system in Simulink andachieved their objective of reducing the fuel consumption
2 Journal of Control Science and Engineering
Torque (119879)
119879peak
119879cont
Quadrant II (generating)Speed minus
Torque +
Quadrant I (motoring)Speed +
Torque +
Quadrant III (motoring)Speed minus
Torque minus
Quadrant IV (generating)Speed +
Torque minus
119908max
Speed
Figure 1 Torque versus speed motoring and generating quadrants(adopted from [2])
Steinmaurer and del Re [4] used an ISG with an HEV witha statistics-based control to achieve task optimization in anattempt to emulate dynamic programming
Two control algorithms are most commonly investigatedin the literature [1] Kessels et al [5] surveyed the controlalgorithms for HEVs up to 2007 Liu and Peng [6] studiedthe control of Toyota PSHEV using two control algorithmsstochastic dynamic programming and equivalent consump-tions minimization strategies (ECMS) [7] They concludedthat with ECMS frequent shifting should be avoided byadding extra constraints between gear switching decisionswhile with stochastic dynamic programming approach (SDP)an extra input operating gear mode is needed beside theengine speed and the SOC Syed et al [8] used a fuzzy logicgain scheduling algorithm with proportional-integral (PI)controllers which are used with power-split HEV (PSHEV)The results of testing the controller on a Ford Escape showedthat a minimum of four rules are needed to ensure smoothengine speed Canova et al [9] studied the engine startstopdynamics which led to an engine model that is used in alinear quadratic regulator control algorithm Lin et al [10]studied the dynamics of the Toyota Prius PHEV and devel-oped an optimal control energy management strategy andartificial neural networks that was modified to a suboptimalcontroller
Gokasan et al [11] used sliding mode control to limit theseries HEVs to the optimal efficiency operating range andcompared it to PI controllers They used two simultaneoussliding mode controllers one to control the engine speed andthe other to control the engine torque Moura et al [12] usedSDP in PHEV powermanagement and explored the potentialof charge depletion overcharging and the economy of fuel-electric usage in hybrid technology Their approach allowsfor energy management control without prior knowledgeof the driving cycle Johannesson et al [13] developed aGPS-based SDP control algorithm for HEV that utilizes thetravel and location information of the HEV to manage itspower sources to assess the potential benefit of predictivecontrolsThey compared a position-dependent controller anda position-invariant one to an optimal control algorithm
Deceleratingdownhill Braking or decelerating
Load movingunder gravity
Engine is idle
Figure 2 Regeneration opportunities
which showed a slight reduction in fuel consumption whenthe position information is taken into account However theyalso discovered that only the average travel information isneeded and the high level of detail of travel conditions is notneeded
Huang et al [14] used a threshold logic control strategythat is optimized continuously during operation to controla PHEV Lee et al [15] developed neuro-fuzzy techniqueto analyze and learn GPS data and control the hybridvehicle to improve fuel economy and reduce emissions Fu[16] developed a methodology for real-time prediction oftorque availability in interior permanent magnet motorsWhen implemented on HEVs it prevents torque requestsat times when there are no availabilities hence protectingthe battery from depletion According to Borhan et al [17]dynamic programming requires knowledge of the upcomingcycle while ECMS are less sensitive to computation short-sighted and sensitive to their parameters Borhan et al [17]linearized the models and simulated over different drivingcycles They modeled the system using the resistance torquerate of SOC equations and Willanrsquos line method For thisreason MPC strategy for PSHEV has been developed tominimize fuel consumption reduce service brake usage andprevent overcharge and discharge of the battery Hybridpowertrain technology has been studied and implemented inpassenger vehicles and trucks yet very little literature can befound about hybrids in construction equipment applicationsGrammatico et al [18] implemented PSHEV technology insmall wheel loaders and used a classic proportional-integral-derivative controller (PID) to control the system
Themachine investigated in this work is a medium wheelloader (MWL) However the procedures used are generalsuch that they are applicable to other types of workmachines
2 Medium Wheel Loaders andHybrid Technology
The prototype machine selected is a medium wheel loader(MWL) A detailed dynamic machine model is describedin Figure 3 For this purpose accurate virtual models withrelatively high fidelity are needed to attempt mimicking theactual machine However replicating every aspect of the
Journal of Control Science and Engineering 3
machine in virtual reality is impossible due to difficulties inmodeling actual physics A detailed virtual dynamic machinemodel is developed using the mathematical equation andembedded in S-functions in Simulink The model is that isvalidated by comparing it to experimental data There arefour general work cycles for this type of machine truck-loading load-and-carry pile dressing and roading Both thetruck-loading and the load-and-carry are cycles that involvedigging moving earth from one location to another anddumping it at the new location The truck-loading is loadinga truck or a hopper with earth and is categorized into twomajor cycles aggressive truck-loading (ATL) and moderatetruck-loading (MTL) ATL is characterized by its speed andthe machine is operated with the engine at full throttle at alltimes
Moderate truck-loading (MTL) can be as fast as or slowerthan ATL but the operator varies engine speed commandandmay never reach full throttleThe load-and-carry cycle ismoving dirt to a hopper far away from the pile and is definedby the travel distance into short and long Pile dressing ismoving the earth in the pile around to make it ready for thefollowing operations Roading is driving the loader from onelocation to another without being involved in any of loadmovements There are five main systems in loaders namely
(1) engine(2) transmission and lower powertrain(3) hydraulics (for work tool and steering)(4) controllers and(5) chassis (frame cabin and linkages)
The engine is the primary source of energy for thismachine from which all power requirements are drawnThe torque provided by the engine is split between twothe powertrains which are responsible for the motion ofthe loader and the hydraulics which are responsible formainly linkages operations and possibly braking cooling fanmotoring and steering
21 Engine The engine is a four-stroke turbocharged aftercooled injection approximately 9-liter diesel engine withaverage power rating (Figure 4) of 224ndash261 kW at 1800ndash2200 rpm and compression ratio of 17 1 The output grosspower from the engine is not the actual available power to thewheel loader main systems due to the power consumptionsfrom accessories such as the alternator the muffler theemission control and the cooling system [19] Thus thepower losses in the range of 6ndash14must be considered whencalculating the net available power to the powertrain andthe hydraulics Unlike automotive engines the diesel enginespeed in construction equipment applications is limited toabout 2300 rpmThis is due to the need for higher torque val-ues at lower machine speeds the desire for longer engine lifeand reduced fuel consumption The engine dynamic modelallows for the calculation of the torque and speed along thelug curve and calculating the engine fuel consumption via thebrake fuel consumption map With hybrid implementationthis engine is downsized to approximately 7 liters of diesel
Table 1 Transmission ratios
Gear shift Gear ratio4F minus077
3F minus136
2F minus244
1F minus466
N 01R 4222R 2213R 1234R 070
allowing the engine to run in a more efficient zone andmaking the hybrid system more cost effective
22 Transmission and Lower Powertrain The main functionof the powertrain is to transfer the power from the engineto the wheels in order to create the necessary rimpull forthe motion through a series of speed reductions and torquemultiplications The powertrain of a wheel loader consists ofwheels axle reductions differentials axles a torque converter(TC) and transmission (Figure 5 (httpwwwcaranddrivercomphotos-09q43095872010-bmw-activehybrid-x6-auto-matic-transmission-and-electric-motor-diagram-photo-310275))
A TC is a hydrodynamic coupling device which transferspower between its input and output while absorbing thedifference in speed between the two shafts by slipping Thedifference in power between input and output is absorbed anddissipated in the oil inside the TC in the form of heat Thetwo main defining characteristics of the TC are the primarytorque119879119901 which is the value of the rated torque of the impellerat the rated speed 119873rated (usually 1700 rpm) and the torqueratio 119879119903 which is the ratio between the output turbine torqueand the input impeller torqueAlso119879119903 is function of the speedratio 119878119903 which is the ratio between the output turbine speedand the input impeller speed (119873119894) Through the knowledgeof 119879119901 119878119903 and the rated and the input speed119873119894 and the inputtorque 119879119894 can be calculated at various speeds using (1) andFigure 6
119879119894 = 119879119901
1003816
1003816
1003816
1003816
1003816119878119903
(
119873119894
119873rated)
2
(1)
The TC is followed by a planetary gearbox which reducesthe output speed to a desired range for the machine groundvelocity The gearbox is constituted of several planetary geartrains connected together via electrohydraulically controlledbrakes and clutches which are used to select the gear Theexplanation of how the mechanism works is available in theliterature [19]The regular gear train in wheel loaders has fourforward speeds four reverse and one neutral (Table 1) Thedynamic transmission model is based on multiple look-uptables for the gear combinations and clutchesrsquo pressures
23 Hydraulics The hydraulic system of the prototype MWLmachine used in this study is a closed center load sensing
4 Journal of Control Science and Engineering
ECM
BrakeHydraulicPneumaticPneumatic + hydraulic
ECM
Power source
Powertrain
(diesel engine)
ECM
Torque converter +mechanical transmission
Hydrostatic driveTires
TracksMachine traction
and speed
ECM
SteeringHydraulic
ECM
ImplementsHydraulicGround engaging tool (GET)
Tool forceand speed
Cooling fan Radiator
Figure 3 Wheel loader systems
140
160
180
200
220
240
260
280
Pow
er (k
W)
Engine speed (rpm)1200 1400 1600 1800 2000 2200
900
1000
1100
1200
1300
1400
1500
1600
Rating B
Rating C
Rating A
Torq
ue (N
middotm)
Figure 4 Rated engine lug curve and power rating
hydraulics system for the work tool (implement) circuit Thissystem is the common trend in MWL as it avoids dissipationof energy and adapts the amount of flow provided by thepump to meet the needs of the machine thus minimizing thelosses unlike open center systems The load sensing systemcompares the pump pressure at the output to the cylinderspressures and then it adjusts the pumprsquos swash plate basedon that feedback in order to maintain a certain pressuredifferential (Figure 7 (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)) A shuttle valve sensesthe highest pressure between the tilt and lift pressures and
sends it through feedback for comparison Equation (2)governs the angle of the pumprsquos swash plate [2]
120579cmd = 120579offset + 119870 (Δ119875cmd minus (119875119904 minus 119875119871)) (2)
where 120579cmd is the commanded swash plate angle 120579offset isthe swash plate angle offset from the desired position 119870 is againΔ119875cmd is the desired pressure differential119875119904 is the pumpoutlet pressure and 119875119871 is the load pressure
A pressure limiting control system in the pump is calledthe high pressure cut-off or pressure compensator whosefunction is to limit themaximumsystempressure by reducingpump displacement to zero when the set pressure is reachedThe post compensation regulates the amount of flow sentto the tilt and lift based on the pressure feedback fromboth valves such that the valve with the higher pressurereceivesmore flow than that with the lower pressure that is itmaintains a constant pressure drop across the valveThe spoolmovement of a compensator valve is governed by (3)
Δ119909119904 =1
119870spring(119901119904119860 119904 minus 119901119897119860 119897) (3)
where 119901119904 is the supply pressure 119860 119904 is the area of the piston atthe supply pressure side 119901119897 is the load pressure119860 119897 is the areaof the piston at the load pressure side 119870spring is the springcoefficient and Δ119909119904 is the length of the opening from thesupply to the load in the valve Equations (4) through (7)
Journal of Control Science and Engineering 5
1 2
3
4
Figure 5 Transmission
4
2
0
119879119903 t
orqu
e rat
io
400
200
0
119879119903 torque ratio
119879119901 primary torque
0 01 02 03 04 05 06 07 08 09 1119878119903 speed ratio
119879119901
prim
ary
torq
ue (N
middotm)
Figure 6 Torque converter characteristics
govern the dynamics and static hydraulics and the hydro-mechanical equations
119876 =
119881
119875
120573
(4)
119876 = 119862119889119860 radic2Δ119875
120588
(5)
119876 = 119860 (6)
119865 = 119875119860 (7)
where 120573 is the bulk modulus119876 is the volumetric flow rate 119875is the pressure 120588 is the oil density 119862119889 is the valve coefficient119860 is the valve opening area and 119909 is the displacement of thepiston
24 Chassis The chassis is the body and linkages of themachine and it is governed by the mechanical kinematicconstraints and dynamics which can be represented usingmultibody dynamics approaches
25 Electric Hybrid The electric hybrid concept (Figure 8)is used to convert the regenerative mechanical energy toelectrical energy via an electric generator and then store itin electrochemical batteries The electric hybrid consists of
a dual-function motorgenerator (ISG) actuator an inverterthat requires a separate cooling system and a lithium-Ionbattery pack The Li-I battery was selected as it is the mostpromising rechargeable battery technology available and iswidely used in electric hybrid technology [20] The SOCchanges over time interval 119889119905 with discharging or chargingcurrent 119894 The battery SOC can be calculated using (8)
SOC = SOC0 minus int119894 (119905)
119876 (119894 (119905))
119889119905 (8)
where 119876(119894(119905)) is the ampere-hour capacity of the batteryat current rate 119894(119905) While the electric hybrid is the mostexpensive it is receiving the most investment in all mobileindustries due to its maturity With the engine downsizingthe ISG system costs come to neutralThe diesel engine is theprimary power plant electric hybrid is the energy bumperThe ISG adds power to powertrain from the battery whenpower assist is needed It also stores the energy to the batteryfrom the powertrain when a power storage opportunityexists The electric hybrid cooling cycle passes through allthe components where the coolant fluid goes through theshunt tank into radiator through the pump inverter batteryand ISG and then back to radiator Also the shunt tank isconnected to all the hybrid parts to allow air bubbles to escapeand to the radiator to allow coolant to expand and supplyexcess fluid
3 Control Strategies
In this section the proposed control strategy is presentedThecontrol strategy under investigation is a rule-based controlBased on the analysis and the results it will be determinedif gain scheduling via GPS tracking and cycle recognitionalgorithm is needed The rule-based control gains are tunedto different cycles to minimize the fuel consumption bringthe final SOC closest to the initial SOC minimize the cycletime and regulate the minimum engine speed to be closeto 1000 rpm or higher (Figure 9) The high-level Simulinkdiagram is shown (Figure 10)
31 Cycle Model The cycle model is built to mimic thecommands of the real operator sent to the different machinesystems as well as the work area In total there are fourdifferent cycles and hence four different operator modelsused in this investigation ATL MTL short load-and-carry(SLC) and long load-and-carry (LLC) Table 2 shows theinput and output signals to the operator model The cyclemodel is modeled as ldquoif statementsrdquo with multiple possibleoutput scenarios based on time and distance along the cycleEach if statement output corresponds to a group of time-and distance-based operator commands to the machine Thesequence of these commands will result in the machineoperations that will lead to completing the cycle Table 3shows the input and output signals to the controller
32 Rule-Based Logic The rule-based control (Figure 10) isdesigned based on knowledge of machine functions andsystem of the machine This control is torque-based control
6 Journal of Control Science and Engineering
Pump control pistonSwash plate
Load senseCylinder
Bias spring Pump piston
Pump control valve
Compensator
Implement control valve
Figure 7 Load sensing axial piston pump with swash plate (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)
Battery
Power electronics
Engine Motorgenerator Transmission Differential
Figure 8 Electric hybrid concept (adopted from [6])
GPS signal
Path identification Cycle identification
Control logicControlsoftware
block
On-machine sensors Hybrid systemactuators
Figure 9 Rule-based control block diagram
that will send out desired torque to the ISG and based onthis the generator-motor actionwill be determined To decidewhether to charge or assist multiple factors are consideredThese factors are represented by gains and thresholds that willenable reaching the decision These gains and thresholds aretuned for different cycles to achieve optimum robust resultsbased on the criteria listed earlierThegains tuned in this logicare delay timer assist threshold charge threshold low SOCthreshold idle charge torque ISG torque-required thresholdengine speed factor assistcharge threshold offset and theidle charge SOC threshold
Table 2 Operator model signals
Cycle model input signals Cycle model output signals(1) Engine speed (rpm) (1) Desired engine speed (rpm)(2) Desired gear number (2) Gear command(3) Lift displacement (m) (3) Lift command(4) Tilt displacement (m) (4) Tilt command(5) Machine ground velocity (5) Brake command
(ms) (6) Bucket load(7) Grade(8) Drawbar
Table 3 Rule-based logic signals
Input signals Output signals(1) Engine speed (rpm) (1) ISG torque command(2) Desired engine speed (rpm)(3) Engine load torque (Nm)(4) Engine maximum torque (Nm)(5) SOC(6) Gear command(7) Grade
To reach this decision the first step is to multiply theengine torque limit by three of the gains and compare theresults to the engine load torque These gains are the assistthreshold charge threshold and hysteresis factor Anothergain existing to assist in decision making is the low SOCthreshold which is biased to charging when the SOC dropsbelow it For low SOC mode the more aggressive (chargebias) assistcharge thresholds should be arrived at by takingthe standard values and offsetting them by assistchargethreshold offset Along with comparison of the enginespeed and torque demand with the actual this procedurewill determine the torque demand out of the engine Thisstep determines if the engine is capable of supplying thedemanded power on its own needs assistance from thehybrid or has excess power to be used for charging If theengine is idle and the SOC is less than the idle SOC charge
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
2 Journal of Control Science and Engineering
Torque (119879)
119879peak
119879cont
Quadrant II (generating)Speed minus
Torque +
Quadrant I (motoring)Speed +
Torque +
Quadrant III (motoring)Speed minus
Torque minus
Quadrant IV (generating)Speed +
Torque minus
119908max
Speed
Figure 1 Torque versus speed motoring and generating quadrants(adopted from [2])
Steinmaurer and del Re [4] used an ISG with an HEV witha statistics-based control to achieve task optimization in anattempt to emulate dynamic programming
Two control algorithms are most commonly investigatedin the literature [1] Kessels et al [5] surveyed the controlalgorithms for HEVs up to 2007 Liu and Peng [6] studiedthe control of Toyota PSHEV using two control algorithmsstochastic dynamic programming and equivalent consump-tions minimization strategies (ECMS) [7] They concludedthat with ECMS frequent shifting should be avoided byadding extra constraints between gear switching decisionswhile with stochastic dynamic programming approach (SDP)an extra input operating gear mode is needed beside theengine speed and the SOC Syed et al [8] used a fuzzy logicgain scheduling algorithm with proportional-integral (PI)controllers which are used with power-split HEV (PSHEV)The results of testing the controller on a Ford Escape showedthat a minimum of four rules are needed to ensure smoothengine speed Canova et al [9] studied the engine startstopdynamics which led to an engine model that is used in alinear quadratic regulator control algorithm Lin et al [10]studied the dynamics of the Toyota Prius PHEV and devel-oped an optimal control energy management strategy andartificial neural networks that was modified to a suboptimalcontroller
Gokasan et al [11] used sliding mode control to limit theseries HEVs to the optimal efficiency operating range andcompared it to PI controllers They used two simultaneoussliding mode controllers one to control the engine speed andthe other to control the engine torque Moura et al [12] usedSDP in PHEV powermanagement and explored the potentialof charge depletion overcharging and the economy of fuel-electric usage in hybrid technology Their approach allowsfor energy management control without prior knowledgeof the driving cycle Johannesson et al [13] developed aGPS-based SDP control algorithm for HEV that utilizes thetravel and location information of the HEV to manage itspower sources to assess the potential benefit of predictivecontrolsThey compared a position-dependent controller anda position-invariant one to an optimal control algorithm
Deceleratingdownhill Braking or decelerating
Load movingunder gravity
Engine is idle
Figure 2 Regeneration opportunities
which showed a slight reduction in fuel consumption whenthe position information is taken into account However theyalso discovered that only the average travel information isneeded and the high level of detail of travel conditions is notneeded
Huang et al [14] used a threshold logic control strategythat is optimized continuously during operation to controla PHEV Lee et al [15] developed neuro-fuzzy techniqueto analyze and learn GPS data and control the hybridvehicle to improve fuel economy and reduce emissions Fu[16] developed a methodology for real-time prediction oftorque availability in interior permanent magnet motorsWhen implemented on HEVs it prevents torque requestsat times when there are no availabilities hence protectingthe battery from depletion According to Borhan et al [17]dynamic programming requires knowledge of the upcomingcycle while ECMS are less sensitive to computation short-sighted and sensitive to their parameters Borhan et al [17]linearized the models and simulated over different drivingcycles They modeled the system using the resistance torquerate of SOC equations and Willanrsquos line method For thisreason MPC strategy for PSHEV has been developed tominimize fuel consumption reduce service brake usage andprevent overcharge and discharge of the battery Hybridpowertrain technology has been studied and implemented inpassenger vehicles and trucks yet very little literature can befound about hybrids in construction equipment applicationsGrammatico et al [18] implemented PSHEV technology insmall wheel loaders and used a classic proportional-integral-derivative controller (PID) to control the system
Themachine investigated in this work is a medium wheelloader (MWL) However the procedures used are generalsuch that they are applicable to other types of workmachines
2 Medium Wheel Loaders andHybrid Technology
The prototype machine selected is a medium wheel loader(MWL) A detailed dynamic machine model is describedin Figure 3 For this purpose accurate virtual models withrelatively high fidelity are needed to attempt mimicking theactual machine However replicating every aspect of the
Journal of Control Science and Engineering 3
machine in virtual reality is impossible due to difficulties inmodeling actual physics A detailed virtual dynamic machinemodel is developed using the mathematical equation andembedded in S-functions in Simulink The model is that isvalidated by comparing it to experimental data There arefour general work cycles for this type of machine truck-loading load-and-carry pile dressing and roading Both thetruck-loading and the load-and-carry are cycles that involvedigging moving earth from one location to another anddumping it at the new location The truck-loading is loadinga truck or a hopper with earth and is categorized into twomajor cycles aggressive truck-loading (ATL) and moderatetruck-loading (MTL) ATL is characterized by its speed andthe machine is operated with the engine at full throttle at alltimes
Moderate truck-loading (MTL) can be as fast as or slowerthan ATL but the operator varies engine speed commandandmay never reach full throttleThe load-and-carry cycle ismoving dirt to a hopper far away from the pile and is definedby the travel distance into short and long Pile dressing ismoving the earth in the pile around to make it ready for thefollowing operations Roading is driving the loader from onelocation to another without being involved in any of loadmovements There are five main systems in loaders namely
(1) engine(2) transmission and lower powertrain(3) hydraulics (for work tool and steering)(4) controllers and(5) chassis (frame cabin and linkages)
The engine is the primary source of energy for thismachine from which all power requirements are drawnThe torque provided by the engine is split between twothe powertrains which are responsible for the motion ofthe loader and the hydraulics which are responsible formainly linkages operations and possibly braking cooling fanmotoring and steering
21 Engine The engine is a four-stroke turbocharged aftercooled injection approximately 9-liter diesel engine withaverage power rating (Figure 4) of 224ndash261 kW at 1800ndash2200 rpm and compression ratio of 17 1 The output grosspower from the engine is not the actual available power to thewheel loader main systems due to the power consumptionsfrom accessories such as the alternator the muffler theemission control and the cooling system [19] Thus thepower losses in the range of 6ndash14must be considered whencalculating the net available power to the powertrain andthe hydraulics Unlike automotive engines the diesel enginespeed in construction equipment applications is limited toabout 2300 rpmThis is due to the need for higher torque val-ues at lower machine speeds the desire for longer engine lifeand reduced fuel consumption The engine dynamic modelallows for the calculation of the torque and speed along thelug curve and calculating the engine fuel consumption via thebrake fuel consumption map With hybrid implementationthis engine is downsized to approximately 7 liters of diesel
Table 1 Transmission ratios
Gear shift Gear ratio4F minus077
3F minus136
2F minus244
1F minus466
N 01R 4222R 2213R 1234R 070
allowing the engine to run in a more efficient zone andmaking the hybrid system more cost effective
22 Transmission and Lower Powertrain The main functionof the powertrain is to transfer the power from the engineto the wheels in order to create the necessary rimpull forthe motion through a series of speed reductions and torquemultiplications The powertrain of a wheel loader consists ofwheels axle reductions differentials axles a torque converter(TC) and transmission (Figure 5 (httpwwwcaranddrivercomphotos-09q43095872010-bmw-activehybrid-x6-auto-matic-transmission-and-electric-motor-diagram-photo-310275))
A TC is a hydrodynamic coupling device which transferspower between its input and output while absorbing thedifference in speed between the two shafts by slipping Thedifference in power between input and output is absorbed anddissipated in the oil inside the TC in the form of heat Thetwo main defining characteristics of the TC are the primarytorque119879119901 which is the value of the rated torque of the impellerat the rated speed 119873rated (usually 1700 rpm) and the torqueratio 119879119903 which is the ratio between the output turbine torqueand the input impeller torqueAlso119879119903 is function of the speedratio 119878119903 which is the ratio between the output turbine speedand the input impeller speed (119873119894) Through the knowledgeof 119879119901 119878119903 and the rated and the input speed119873119894 and the inputtorque 119879119894 can be calculated at various speeds using (1) andFigure 6
119879119894 = 119879119901
1003816
1003816
1003816
1003816
1003816119878119903
(
119873119894
119873rated)
2
(1)
The TC is followed by a planetary gearbox which reducesthe output speed to a desired range for the machine groundvelocity The gearbox is constituted of several planetary geartrains connected together via electrohydraulically controlledbrakes and clutches which are used to select the gear Theexplanation of how the mechanism works is available in theliterature [19]The regular gear train in wheel loaders has fourforward speeds four reverse and one neutral (Table 1) Thedynamic transmission model is based on multiple look-uptables for the gear combinations and clutchesrsquo pressures
23 Hydraulics The hydraulic system of the prototype MWLmachine used in this study is a closed center load sensing
4 Journal of Control Science and Engineering
ECM
BrakeHydraulicPneumaticPneumatic + hydraulic
ECM
Power source
Powertrain
(diesel engine)
ECM
Torque converter +mechanical transmission
Hydrostatic driveTires
TracksMachine traction
and speed
ECM
SteeringHydraulic
ECM
ImplementsHydraulicGround engaging tool (GET)
Tool forceand speed
Cooling fan Radiator
Figure 3 Wheel loader systems
140
160
180
200
220
240
260
280
Pow
er (k
W)
Engine speed (rpm)1200 1400 1600 1800 2000 2200
900
1000
1100
1200
1300
1400
1500
1600
Rating B
Rating C
Rating A
Torq
ue (N
middotm)
Figure 4 Rated engine lug curve and power rating
hydraulics system for the work tool (implement) circuit Thissystem is the common trend in MWL as it avoids dissipationof energy and adapts the amount of flow provided by thepump to meet the needs of the machine thus minimizing thelosses unlike open center systems The load sensing systemcompares the pump pressure at the output to the cylinderspressures and then it adjusts the pumprsquos swash plate basedon that feedback in order to maintain a certain pressuredifferential (Figure 7 (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)) A shuttle valve sensesthe highest pressure between the tilt and lift pressures and
sends it through feedback for comparison Equation (2)governs the angle of the pumprsquos swash plate [2]
120579cmd = 120579offset + 119870 (Δ119875cmd minus (119875119904 minus 119875119871)) (2)
where 120579cmd is the commanded swash plate angle 120579offset isthe swash plate angle offset from the desired position 119870 is againΔ119875cmd is the desired pressure differential119875119904 is the pumpoutlet pressure and 119875119871 is the load pressure
A pressure limiting control system in the pump is calledthe high pressure cut-off or pressure compensator whosefunction is to limit themaximumsystempressure by reducingpump displacement to zero when the set pressure is reachedThe post compensation regulates the amount of flow sentto the tilt and lift based on the pressure feedback fromboth valves such that the valve with the higher pressurereceivesmore flow than that with the lower pressure that is itmaintains a constant pressure drop across the valveThe spoolmovement of a compensator valve is governed by (3)
Δ119909119904 =1
119870spring(119901119904119860 119904 minus 119901119897119860 119897) (3)
where 119901119904 is the supply pressure 119860 119904 is the area of the piston atthe supply pressure side 119901119897 is the load pressure119860 119897 is the areaof the piston at the load pressure side 119870spring is the springcoefficient and Δ119909119904 is the length of the opening from thesupply to the load in the valve Equations (4) through (7)
Journal of Control Science and Engineering 5
1 2
3
4
Figure 5 Transmission
4
2
0
119879119903 t
orqu
e rat
io
400
200
0
119879119903 torque ratio
119879119901 primary torque
0 01 02 03 04 05 06 07 08 09 1119878119903 speed ratio
119879119901
prim
ary
torq
ue (N
middotm)
Figure 6 Torque converter characteristics
govern the dynamics and static hydraulics and the hydro-mechanical equations
119876 =
119881
119875
120573
(4)
119876 = 119862119889119860 radic2Δ119875
120588
(5)
119876 = 119860 (6)
119865 = 119875119860 (7)
where 120573 is the bulk modulus119876 is the volumetric flow rate 119875is the pressure 120588 is the oil density 119862119889 is the valve coefficient119860 is the valve opening area and 119909 is the displacement of thepiston
24 Chassis The chassis is the body and linkages of themachine and it is governed by the mechanical kinematicconstraints and dynamics which can be represented usingmultibody dynamics approaches
25 Electric Hybrid The electric hybrid concept (Figure 8)is used to convert the regenerative mechanical energy toelectrical energy via an electric generator and then store itin electrochemical batteries The electric hybrid consists of
a dual-function motorgenerator (ISG) actuator an inverterthat requires a separate cooling system and a lithium-Ionbattery pack The Li-I battery was selected as it is the mostpromising rechargeable battery technology available and iswidely used in electric hybrid technology [20] The SOCchanges over time interval 119889119905 with discharging or chargingcurrent 119894 The battery SOC can be calculated using (8)
SOC = SOC0 minus int119894 (119905)
119876 (119894 (119905))
119889119905 (8)
where 119876(119894(119905)) is the ampere-hour capacity of the batteryat current rate 119894(119905) While the electric hybrid is the mostexpensive it is receiving the most investment in all mobileindustries due to its maturity With the engine downsizingthe ISG system costs come to neutralThe diesel engine is theprimary power plant electric hybrid is the energy bumperThe ISG adds power to powertrain from the battery whenpower assist is needed It also stores the energy to the batteryfrom the powertrain when a power storage opportunityexists The electric hybrid cooling cycle passes through allthe components where the coolant fluid goes through theshunt tank into radiator through the pump inverter batteryand ISG and then back to radiator Also the shunt tank isconnected to all the hybrid parts to allow air bubbles to escapeand to the radiator to allow coolant to expand and supplyexcess fluid
3 Control Strategies
In this section the proposed control strategy is presentedThecontrol strategy under investigation is a rule-based controlBased on the analysis and the results it will be determinedif gain scheduling via GPS tracking and cycle recognitionalgorithm is needed The rule-based control gains are tunedto different cycles to minimize the fuel consumption bringthe final SOC closest to the initial SOC minimize the cycletime and regulate the minimum engine speed to be closeto 1000 rpm or higher (Figure 9) The high-level Simulinkdiagram is shown (Figure 10)
31 Cycle Model The cycle model is built to mimic thecommands of the real operator sent to the different machinesystems as well as the work area In total there are fourdifferent cycles and hence four different operator modelsused in this investigation ATL MTL short load-and-carry(SLC) and long load-and-carry (LLC) Table 2 shows theinput and output signals to the operator model The cyclemodel is modeled as ldquoif statementsrdquo with multiple possibleoutput scenarios based on time and distance along the cycleEach if statement output corresponds to a group of time-and distance-based operator commands to the machine Thesequence of these commands will result in the machineoperations that will lead to completing the cycle Table 3shows the input and output signals to the controller
32 Rule-Based Logic The rule-based control (Figure 10) isdesigned based on knowledge of machine functions andsystem of the machine This control is torque-based control
6 Journal of Control Science and Engineering
Pump control pistonSwash plate
Load senseCylinder
Bias spring Pump piston
Pump control valve
Compensator
Implement control valve
Figure 7 Load sensing axial piston pump with swash plate (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)
Battery
Power electronics
Engine Motorgenerator Transmission Differential
Figure 8 Electric hybrid concept (adopted from [6])
GPS signal
Path identification Cycle identification
Control logicControlsoftware
block
On-machine sensors Hybrid systemactuators
Figure 9 Rule-based control block diagram
that will send out desired torque to the ISG and based onthis the generator-motor actionwill be determined To decidewhether to charge or assist multiple factors are consideredThese factors are represented by gains and thresholds that willenable reaching the decision These gains and thresholds aretuned for different cycles to achieve optimum robust resultsbased on the criteria listed earlierThegains tuned in this logicare delay timer assist threshold charge threshold low SOCthreshold idle charge torque ISG torque-required thresholdengine speed factor assistcharge threshold offset and theidle charge SOC threshold
Table 2 Operator model signals
Cycle model input signals Cycle model output signals(1) Engine speed (rpm) (1) Desired engine speed (rpm)(2) Desired gear number (2) Gear command(3) Lift displacement (m) (3) Lift command(4) Tilt displacement (m) (4) Tilt command(5) Machine ground velocity (5) Brake command
(ms) (6) Bucket load(7) Grade(8) Drawbar
Table 3 Rule-based logic signals
Input signals Output signals(1) Engine speed (rpm) (1) ISG torque command(2) Desired engine speed (rpm)(3) Engine load torque (Nm)(4) Engine maximum torque (Nm)(5) SOC(6) Gear command(7) Grade
To reach this decision the first step is to multiply theengine torque limit by three of the gains and compare theresults to the engine load torque These gains are the assistthreshold charge threshold and hysteresis factor Anothergain existing to assist in decision making is the low SOCthreshold which is biased to charging when the SOC dropsbelow it For low SOC mode the more aggressive (chargebias) assistcharge thresholds should be arrived at by takingthe standard values and offsetting them by assistchargethreshold offset Along with comparison of the enginespeed and torque demand with the actual this procedurewill determine the torque demand out of the engine Thisstep determines if the engine is capable of supplying thedemanded power on its own needs assistance from thehybrid or has excess power to be used for charging If theengine is idle and the SOC is less than the idle SOC charge
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Control Science and Engineering 3
machine in virtual reality is impossible due to difficulties inmodeling actual physics A detailed virtual dynamic machinemodel is developed using the mathematical equation andembedded in S-functions in Simulink The model is that isvalidated by comparing it to experimental data There arefour general work cycles for this type of machine truck-loading load-and-carry pile dressing and roading Both thetruck-loading and the load-and-carry are cycles that involvedigging moving earth from one location to another anddumping it at the new location The truck-loading is loadinga truck or a hopper with earth and is categorized into twomajor cycles aggressive truck-loading (ATL) and moderatetruck-loading (MTL) ATL is characterized by its speed andthe machine is operated with the engine at full throttle at alltimes
Moderate truck-loading (MTL) can be as fast as or slowerthan ATL but the operator varies engine speed commandandmay never reach full throttleThe load-and-carry cycle ismoving dirt to a hopper far away from the pile and is definedby the travel distance into short and long Pile dressing ismoving the earth in the pile around to make it ready for thefollowing operations Roading is driving the loader from onelocation to another without being involved in any of loadmovements There are five main systems in loaders namely
(1) engine(2) transmission and lower powertrain(3) hydraulics (for work tool and steering)(4) controllers and(5) chassis (frame cabin and linkages)
The engine is the primary source of energy for thismachine from which all power requirements are drawnThe torque provided by the engine is split between twothe powertrains which are responsible for the motion ofthe loader and the hydraulics which are responsible formainly linkages operations and possibly braking cooling fanmotoring and steering
21 Engine The engine is a four-stroke turbocharged aftercooled injection approximately 9-liter diesel engine withaverage power rating (Figure 4) of 224ndash261 kW at 1800ndash2200 rpm and compression ratio of 17 1 The output grosspower from the engine is not the actual available power to thewheel loader main systems due to the power consumptionsfrom accessories such as the alternator the muffler theemission control and the cooling system [19] Thus thepower losses in the range of 6ndash14must be considered whencalculating the net available power to the powertrain andthe hydraulics Unlike automotive engines the diesel enginespeed in construction equipment applications is limited toabout 2300 rpmThis is due to the need for higher torque val-ues at lower machine speeds the desire for longer engine lifeand reduced fuel consumption The engine dynamic modelallows for the calculation of the torque and speed along thelug curve and calculating the engine fuel consumption via thebrake fuel consumption map With hybrid implementationthis engine is downsized to approximately 7 liters of diesel
Table 1 Transmission ratios
Gear shift Gear ratio4F minus077
3F minus136
2F minus244
1F minus466
N 01R 4222R 2213R 1234R 070
allowing the engine to run in a more efficient zone andmaking the hybrid system more cost effective
22 Transmission and Lower Powertrain The main functionof the powertrain is to transfer the power from the engineto the wheels in order to create the necessary rimpull forthe motion through a series of speed reductions and torquemultiplications The powertrain of a wheel loader consists ofwheels axle reductions differentials axles a torque converter(TC) and transmission (Figure 5 (httpwwwcaranddrivercomphotos-09q43095872010-bmw-activehybrid-x6-auto-matic-transmission-and-electric-motor-diagram-photo-310275))
A TC is a hydrodynamic coupling device which transferspower between its input and output while absorbing thedifference in speed between the two shafts by slipping Thedifference in power between input and output is absorbed anddissipated in the oil inside the TC in the form of heat Thetwo main defining characteristics of the TC are the primarytorque119879119901 which is the value of the rated torque of the impellerat the rated speed 119873rated (usually 1700 rpm) and the torqueratio 119879119903 which is the ratio between the output turbine torqueand the input impeller torqueAlso119879119903 is function of the speedratio 119878119903 which is the ratio between the output turbine speedand the input impeller speed (119873119894) Through the knowledgeof 119879119901 119878119903 and the rated and the input speed119873119894 and the inputtorque 119879119894 can be calculated at various speeds using (1) andFigure 6
119879119894 = 119879119901
1003816
1003816
1003816
1003816
1003816119878119903
(
119873119894
119873rated)
2
(1)
The TC is followed by a planetary gearbox which reducesthe output speed to a desired range for the machine groundvelocity The gearbox is constituted of several planetary geartrains connected together via electrohydraulically controlledbrakes and clutches which are used to select the gear Theexplanation of how the mechanism works is available in theliterature [19]The regular gear train in wheel loaders has fourforward speeds four reverse and one neutral (Table 1) Thedynamic transmission model is based on multiple look-uptables for the gear combinations and clutchesrsquo pressures
23 Hydraulics The hydraulic system of the prototype MWLmachine used in this study is a closed center load sensing
4 Journal of Control Science and Engineering
ECM
BrakeHydraulicPneumaticPneumatic + hydraulic
ECM
Power source
Powertrain
(diesel engine)
ECM
Torque converter +mechanical transmission
Hydrostatic driveTires
TracksMachine traction
and speed
ECM
SteeringHydraulic
ECM
ImplementsHydraulicGround engaging tool (GET)
Tool forceand speed
Cooling fan Radiator
Figure 3 Wheel loader systems
140
160
180
200
220
240
260
280
Pow
er (k
W)
Engine speed (rpm)1200 1400 1600 1800 2000 2200
900
1000
1100
1200
1300
1400
1500
1600
Rating B
Rating C
Rating A
Torq
ue (N
middotm)
Figure 4 Rated engine lug curve and power rating
hydraulics system for the work tool (implement) circuit Thissystem is the common trend in MWL as it avoids dissipationof energy and adapts the amount of flow provided by thepump to meet the needs of the machine thus minimizing thelosses unlike open center systems The load sensing systemcompares the pump pressure at the output to the cylinderspressures and then it adjusts the pumprsquos swash plate basedon that feedback in order to maintain a certain pressuredifferential (Figure 7 (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)) A shuttle valve sensesthe highest pressure between the tilt and lift pressures and
sends it through feedback for comparison Equation (2)governs the angle of the pumprsquos swash plate [2]
120579cmd = 120579offset + 119870 (Δ119875cmd minus (119875119904 minus 119875119871)) (2)
where 120579cmd is the commanded swash plate angle 120579offset isthe swash plate angle offset from the desired position 119870 is againΔ119875cmd is the desired pressure differential119875119904 is the pumpoutlet pressure and 119875119871 is the load pressure
A pressure limiting control system in the pump is calledthe high pressure cut-off or pressure compensator whosefunction is to limit themaximumsystempressure by reducingpump displacement to zero when the set pressure is reachedThe post compensation regulates the amount of flow sentto the tilt and lift based on the pressure feedback fromboth valves such that the valve with the higher pressurereceivesmore flow than that with the lower pressure that is itmaintains a constant pressure drop across the valveThe spoolmovement of a compensator valve is governed by (3)
Δ119909119904 =1
119870spring(119901119904119860 119904 minus 119901119897119860 119897) (3)
where 119901119904 is the supply pressure 119860 119904 is the area of the piston atthe supply pressure side 119901119897 is the load pressure119860 119897 is the areaof the piston at the load pressure side 119870spring is the springcoefficient and Δ119909119904 is the length of the opening from thesupply to the load in the valve Equations (4) through (7)
Journal of Control Science and Engineering 5
1 2
3
4
Figure 5 Transmission
4
2
0
119879119903 t
orqu
e rat
io
400
200
0
119879119903 torque ratio
119879119901 primary torque
0 01 02 03 04 05 06 07 08 09 1119878119903 speed ratio
119879119901
prim
ary
torq
ue (N
middotm)
Figure 6 Torque converter characteristics
govern the dynamics and static hydraulics and the hydro-mechanical equations
119876 =
119881
119875
120573
(4)
119876 = 119862119889119860 radic2Δ119875
120588
(5)
119876 = 119860 (6)
119865 = 119875119860 (7)
where 120573 is the bulk modulus119876 is the volumetric flow rate 119875is the pressure 120588 is the oil density 119862119889 is the valve coefficient119860 is the valve opening area and 119909 is the displacement of thepiston
24 Chassis The chassis is the body and linkages of themachine and it is governed by the mechanical kinematicconstraints and dynamics which can be represented usingmultibody dynamics approaches
25 Electric Hybrid The electric hybrid concept (Figure 8)is used to convert the regenerative mechanical energy toelectrical energy via an electric generator and then store itin electrochemical batteries The electric hybrid consists of
a dual-function motorgenerator (ISG) actuator an inverterthat requires a separate cooling system and a lithium-Ionbattery pack The Li-I battery was selected as it is the mostpromising rechargeable battery technology available and iswidely used in electric hybrid technology [20] The SOCchanges over time interval 119889119905 with discharging or chargingcurrent 119894 The battery SOC can be calculated using (8)
SOC = SOC0 minus int119894 (119905)
119876 (119894 (119905))
119889119905 (8)
where 119876(119894(119905)) is the ampere-hour capacity of the batteryat current rate 119894(119905) While the electric hybrid is the mostexpensive it is receiving the most investment in all mobileindustries due to its maturity With the engine downsizingthe ISG system costs come to neutralThe diesel engine is theprimary power plant electric hybrid is the energy bumperThe ISG adds power to powertrain from the battery whenpower assist is needed It also stores the energy to the batteryfrom the powertrain when a power storage opportunityexists The electric hybrid cooling cycle passes through allthe components where the coolant fluid goes through theshunt tank into radiator through the pump inverter batteryand ISG and then back to radiator Also the shunt tank isconnected to all the hybrid parts to allow air bubbles to escapeand to the radiator to allow coolant to expand and supplyexcess fluid
3 Control Strategies
In this section the proposed control strategy is presentedThecontrol strategy under investigation is a rule-based controlBased on the analysis and the results it will be determinedif gain scheduling via GPS tracking and cycle recognitionalgorithm is needed The rule-based control gains are tunedto different cycles to minimize the fuel consumption bringthe final SOC closest to the initial SOC minimize the cycletime and regulate the minimum engine speed to be closeto 1000 rpm or higher (Figure 9) The high-level Simulinkdiagram is shown (Figure 10)
31 Cycle Model The cycle model is built to mimic thecommands of the real operator sent to the different machinesystems as well as the work area In total there are fourdifferent cycles and hence four different operator modelsused in this investigation ATL MTL short load-and-carry(SLC) and long load-and-carry (LLC) Table 2 shows theinput and output signals to the operator model The cyclemodel is modeled as ldquoif statementsrdquo with multiple possibleoutput scenarios based on time and distance along the cycleEach if statement output corresponds to a group of time-and distance-based operator commands to the machine Thesequence of these commands will result in the machineoperations that will lead to completing the cycle Table 3shows the input and output signals to the controller
32 Rule-Based Logic The rule-based control (Figure 10) isdesigned based on knowledge of machine functions andsystem of the machine This control is torque-based control
6 Journal of Control Science and Engineering
Pump control pistonSwash plate
Load senseCylinder
Bias spring Pump piston
Pump control valve
Compensator
Implement control valve
Figure 7 Load sensing axial piston pump with swash plate (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)
Battery
Power electronics
Engine Motorgenerator Transmission Differential
Figure 8 Electric hybrid concept (adopted from [6])
GPS signal
Path identification Cycle identification
Control logicControlsoftware
block
On-machine sensors Hybrid systemactuators
Figure 9 Rule-based control block diagram
that will send out desired torque to the ISG and based onthis the generator-motor actionwill be determined To decidewhether to charge or assist multiple factors are consideredThese factors are represented by gains and thresholds that willenable reaching the decision These gains and thresholds aretuned for different cycles to achieve optimum robust resultsbased on the criteria listed earlierThegains tuned in this logicare delay timer assist threshold charge threshold low SOCthreshold idle charge torque ISG torque-required thresholdengine speed factor assistcharge threshold offset and theidle charge SOC threshold
Table 2 Operator model signals
Cycle model input signals Cycle model output signals(1) Engine speed (rpm) (1) Desired engine speed (rpm)(2) Desired gear number (2) Gear command(3) Lift displacement (m) (3) Lift command(4) Tilt displacement (m) (4) Tilt command(5) Machine ground velocity (5) Brake command
(ms) (6) Bucket load(7) Grade(8) Drawbar
Table 3 Rule-based logic signals
Input signals Output signals(1) Engine speed (rpm) (1) ISG torque command(2) Desired engine speed (rpm)(3) Engine load torque (Nm)(4) Engine maximum torque (Nm)(5) SOC(6) Gear command(7) Grade
To reach this decision the first step is to multiply theengine torque limit by three of the gains and compare theresults to the engine load torque These gains are the assistthreshold charge threshold and hysteresis factor Anothergain existing to assist in decision making is the low SOCthreshold which is biased to charging when the SOC dropsbelow it For low SOC mode the more aggressive (chargebias) assistcharge thresholds should be arrived at by takingthe standard values and offsetting them by assistchargethreshold offset Along with comparison of the enginespeed and torque demand with the actual this procedurewill determine the torque demand out of the engine Thisstep determines if the engine is capable of supplying thedemanded power on its own needs assistance from thehybrid or has excess power to be used for charging If theengine is idle and the SOC is less than the idle SOC charge
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
4 Journal of Control Science and Engineering
ECM
BrakeHydraulicPneumaticPneumatic + hydraulic
ECM
Power source
Powertrain
(diesel engine)
ECM
Torque converter +mechanical transmission
Hydrostatic driveTires
TracksMachine traction
and speed
ECM
SteeringHydraulic
ECM
ImplementsHydraulicGround engaging tool (GET)
Tool forceand speed
Cooling fan Radiator
Figure 3 Wheel loader systems
140
160
180
200
220
240
260
280
Pow
er (k
W)
Engine speed (rpm)1200 1400 1600 1800 2000 2200
900
1000
1100
1200
1300
1400
1500
1600
Rating B
Rating C
Rating A
Torq
ue (N
middotm)
Figure 4 Rated engine lug curve and power rating
hydraulics system for the work tool (implement) circuit Thissystem is the common trend in MWL as it avoids dissipationof energy and adapts the amount of flow provided by thepump to meet the needs of the machine thus minimizing thelosses unlike open center systems The load sensing systemcompares the pump pressure at the output to the cylinderspressures and then it adjusts the pumprsquos swash plate basedon that feedback in order to maintain a certain pressuredifferential (Figure 7 (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)) A shuttle valve sensesthe highest pressure between the tilt and lift pressures and
sends it through feedback for comparison Equation (2)governs the angle of the pumprsquos swash plate [2]
120579cmd = 120579offset + 119870 (Δ119875cmd minus (119875119904 minus 119875119871)) (2)
where 120579cmd is the commanded swash plate angle 120579offset isthe swash plate angle offset from the desired position 119870 is againΔ119875cmd is the desired pressure differential119875119904 is the pumpoutlet pressure and 119875119871 is the load pressure
A pressure limiting control system in the pump is calledthe high pressure cut-off or pressure compensator whosefunction is to limit themaximumsystempressure by reducingpump displacement to zero when the set pressure is reachedThe post compensation regulates the amount of flow sentto the tilt and lift based on the pressure feedback fromboth valves such that the valve with the higher pressurereceivesmore flow than that with the lower pressure that is itmaintains a constant pressure drop across the valveThe spoolmovement of a compensator valve is governed by (3)
Δ119909119904 =1
119870spring(119901119904119860 119904 minus 119901119897119860 119897) (3)
where 119901119904 is the supply pressure 119860 119904 is the area of the piston atthe supply pressure side 119901119897 is the load pressure119860 119897 is the areaof the piston at the load pressure side 119870spring is the springcoefficient and Δ119909119904 is the length of the opening from thesupply to the load in the valve Equations (4) through (7)
Journal of Control Science and Engineering 5
1 2
3
4
Figure 5 Transmission
4
2
0
119879119903 t
orqu
e rat
io
400
200
0
119879119903 torque ratio
119879119901 primary torque
0 01 02 03 04 05 06 07 08 09 1119878119903 speed ratio
119879119901
prim
ary
torq
ue (N
middotm)
Figure 6 Torque converter characteristics
govern the dynamics and static hydraulics and the hydro-mechanical equations
119876 =
119881
119875
120573
(4)
119876 = 119862119889119860 radic2Δ119875
120588
(5)
119876 = 119860 (6)
119865 = 119875119860 (7)
where 120573 is the bulk modulus119876 is the volumetric flow rate 119875is the pressure 120588 is the oil density 119862119889 is the valve coefficient119860 is the valve opening area and 119909 is the displacement of thepiston
24 Chassis The chassis is the body and linkages of themachine and it is governed by the mechanical kinematicconstraints and dynamics which can be represented usingmultibody dynamics approaches
25 Electric Hybrid The electric hybrid concept (Figure 8)is used to convert the regenerative mechanical energy toelectrical energy via an electric generator and then store itin electrochemical batteries The electric hybrid consists of
a dual-function motorgenerator (ISG) actuator an inverterthat requires a separate cooling system and a lithium-Ionbattery pack The Li-I battery was selected as it is the mostpromising rechargeable battery technology available and iswidely used in electric hybrid technology [20] The SOCchanges over time interval 119889119905 with discharging or chargingcurrent 119894 The battery SOC can be calculated using (8)
SOC = SOC0 minus int119894 (119905)
119876 (119894 (119905))
119889119905 (8)
where 119876(119894(119905)) is the ampere-hour capacity of the batteryat current rate 119894(119905) While the electric hybrid is the mostexpensive it is receiving the most investment in all mobileindustries due to its maturity With the engine downsizingthe ISG system costs come to neutralThe diesel engine is theprimary power plant electric hybrid is the energy bumperThe ISG adds power to powertrain from the battery whenpower assist is needed It also stores the energy to the batteryfrom the powertrain when a power storage opportunityexists The electric hybrid cooling cycle passes through allthe components where the coolant fluid goes through theshunt tank into radiator through the pump inverter batteryand ISG and then back to radiator Also the shunt tank isconnected to all the hybrid parts to allow air bubbles to escapeand to the radiator to allow coolant to expand and supplyexcess fluid
3 Control Strategies
In this section the proposed control strategy is presentedThecontrol strategy under investigation is a rule-based controlBased on the analysis and the results it will be determinedif gain scheduling via GPS tracking and cycle recognitionalgorithm is needed The rule-based control gains are tunedto different cycles to minimize the fuel consumption bringthe final SOC closest to the initial SOC minimize the cycletime and regulate the minimum engine speed to be closeto 1000 rpm or higher (Figure 9) The high-level Simulinkdiagram is shown (Figure 10)
31 Cycle Model The cycle model is built to mimic thecommands of the real operator sent to the different machinesystems as well as the work area In total there are fourdifferent cycles and hence four different operator modelsused in this investigation ATL MTL short load-and-carry(SLC) and long load-and-carry (LLC) Table 2 shows theinput and output signals to the operator model The cyclemodel is modeled as ldquoif statementsrdquo with multiple possibleoutput scenarios based on time and distance along the cycleEach if statement output corresponds to a group of time-and distance-based operator commands to the machine Thesequence of these commands will result in the machineoperations that will lead to completing the cycle Table 3shows the input and output signals to the controller
32 Rule-Based Logic The rule-based control (Figure 10) isdesigned based on knowledge of machine functions andsystem of the machine This control is torque-based control
6 Journal of Control Science and Engineering
Pump control pistonSwash plate
Load senseCylinder
Bias spring Pump piston
Pump control valve
Compensator
Implement control valve
Figure 7 Load sensing axial piston pump with swash plate (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)
Battery
Power electronics
Engine Motorgenerator Transmission Differential
Figure 8 Electric hybrid concept (adopted from [6])
GPS signal
Path identification Cycle identification
Control logicControlsoftware
block
On-machine sensors Hybrid systemactuators
Figure 9 Rule-based control block diagram
that will send out desired torque to the ISG and based onthis the generator-motor actionwill be determined To decidewhether to charge or assist multiple factors are consideredThese factors are represented by gains and thresholds that willenable reaching the decision These gains and thresholds aretuned for different cycles to achieve optimum robust resultsbased on the criteria listed earlierThegains tuned in this logicare delay timer assist threshold charge threshold low SOCthreshold idle charge torque ISG torque-required thresholdengine speed factor assistcharge threshold offset and theidle charge SOC threshold
Table 2 Operator model signals
Cycle model input signals Cycle model output signals(1) Engine speed (rpm) (1) Desired engine speed (rpm)(2) Desired gear number (2) Gear command(3) Lift displacement (m) (3) Lift command(4) Tilt displacement (m) (4) Tilt command(5) Machine ground velocity (5) Brake command
(ms) (6) Bucket load(7) Grade(8) Drawbar
Table 3 Rule-based logic signals
Input signals Output signals(1) Engine speed (rpm) (1) ISG torque command(2) Desired engine speed (rpm)(3) Engine load torque (Nm)(4) Engine maximum torque (Nm)(5) SOC(6) Gear command(7) Grade
To reach this decision the first step is to multiply theengine torque limit by three of the gains and compare theresults to the engine load torque These gains are the assistthreshold charge threshold and hysteresis factor Anothergain existing to assist in decision making is the low SOCthreshold which is biased to charging when the SOC dropsbelow it For low SOC mode the more aggressive (chargebias) assistcharge thresholds should be arrived at by takingthe standard values and offsetting them by assistchargethreshold offset Along with comparison of the enginespeed and torque demand with the actual this procedurewill determine the torque demand out of the engine Thisstep determines if the engine is capable of supplying thedemanded power on its own needs assistance from thehybrid or has excess power to be used for charging If theengine is idle and the SOC is less than the idle SOC charge
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Control Science and Engineering 5
1 2
3
4
Figure 5 Transmission
4
2
0
119879119903 t
orqu
e rat
io
400
200
0
119879119903 torque ratio
119879119901 primary torque
0 01 02 03 04 05 06 07 08 09 1119878119903 speed ratio
119879119901
prim
ary
torq
ue (N
middotm)
Figure 6 Torque converter characteristics
govern the dynamics and static hydraulics and the hydro-mechanical equations
119876 =
119881
119875
120573
(4)
119876 = 119862119889119860 radic2Δ119875
120588
(5)
119876 = 119860 (6)
119865 = 119875119860 (7)
where 120573 is the bulk modulus119876 is the volumetric flow rate 119875is the pressure 120588 is the oil density 119862119889 is the valve coefficient119860 is the valve opening area and 119909 is the displacement of thepiston
24 Chassis The chassis is the body and linkages of themachine and it is governed by the mechanical kinematicconstraints and dynamics which can be represented usingmultibody dynamics approaches
25 Electric Hybrid The electric hybrid concept (Figure 8)is used to convert the regenerative mechanical energy toelectrical energy via an electric generator and then store itin electrochemical batteries The electric hybrid consists of
a dual-function motorgenerator (ISG) actuator an inverterthat requires a separate cooling system and a lithium-Ionbattery pack The Li-I battery was selected as it is the mostpromising rechargeable battery technology available and iswidely used in electric hybrid technology [20] The SOCchanges over time interval 119889119905 with discharging or chargingcurrent 119894 The battery SOC can be calculated using (8)
SOC = SOC0 minus int119894 (119905)
119876 (119894 (119905))
119889119905 (8)
where 119876(119894(119905)) is the ampere-hour capacity of the batteryat current rate 119894(119905) While the electric hybrid is the mostexpensive it is receiving the most investment in all mobileindustries due to its maturity With the engine downsizingthe ISG system costs come to neutralThe diesel engine is theprimary power plant electric hybrid is the energy bumperThe ISG adds power to powertrain from the battery whenpower assist is needed It also stores the energy to the batteryfrom the powertrain when a power storage opportunityexists The electric hybrid cooling cycle passes through allthe components where the coolant fluid goes through theshunt tank into radiator through the pump inverter batteryand ISG and then back to radiator Also the shunt tank isconnected to all the hybrid parts to allow air bubbles to escapeand to the radiator to allow coolant to expand and supplyexcess fluid
3 Control Strategies
In this section the proposed control strategy is presentedThecontrol strategy under investigation is a rule-based controlBased on the analysis and the results it will be determinedif gain scheduling via GPS tracking and cycle recognitionalgorithm is needed The rule-based control gains are tunedto different cycles to minimize the fuel consumption bringthe final SOC closest to the initial SOC minimize the cycletime and regulate the minimum engine speed to be closeto 1000 rpm or higher (Figure 9) The high-level Simulinkdiagram is shown (Figure 10)
31 Cycle Model The cycle model is built to mimic thecommands of the real operator sent to the different machinesystems as well as the work area In total there are fourdifferent cycles and hence four different operator modelsused in this investigation ATL MTL short load-and-carry(SLC) and long load-and-carry (LLC) Table 2 shows theinput and output signals to the operator model The cyclemodel is modeled as ldquoif statementsrdquo with multiple possibleoutput scenarios based on time and distance along the cycleEach if statement output corresponds to a group of time-and distance-based operator commands to the machine Thesequence of these commands will result in the machineoperations that will lead to completing the cycle Table 3shows the input and output signals to the controller
32 Rule-Based Logic The rule-based control (Figure 10) isdesigned based on knowledge of machine functions andsystem of the machine This control is torque-based control
6 Journal of Control Science and Engineering
Pump control pistonSwash plate
Load senseCylinder
Bias spring Pump piston
Pump control valve
Compensator
Implement control valve
Figure 7 Load sensing axial piston pump with swash plate (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)
Battery
Power electronics
Engine Motorgenerator Transmission Differential
Figure 8 Electric hybrid concept (adopted from [6])
GPS signal
Path identification Cycle identification
Control logicControlsoftware
block
On-machine sensors Hybrid systemactuators
Figure 9 Rule-based control block diagram
that will send out desired torque to the ISG and based onthis the generator-motor actionwill be determined To decidewhether to charge or assist multiple factors are consideredThese factors are represented by gains and thresholds that willenable reaching the decision These gains and thresholds aretuned for different cycles to achieve optimum robust resultsbased on the criteria listed earlierThegains tuned in this logicare delay timer assist threshold charge threshold low SOCthreshold idle charge torque ISG torque-required thresholdengine speed factor assistcharge threshold offset and theidle charge SOC threshold
Table 2 Operator model signals
Cycle model input signals Cycle model output signals(1) Engine speed (rpm) (1) Desired engine speed (rpm)(2) Desired gear number (2) Gear command(3) Lift displacement (m) (3) Lift command(4) Tilt displacement (m) (4) Tilt command(5) Machine ground velocity (5) Brake command
(ms) (6) Bucket load(7) Grade(8) Drawbar
Table 3 Rule-based logic signals
Input signals Output signals(1) Engine speed (rpm) (1) ISG torque command(2) Desired engine speed (rpm)(3) Engine load torque (Nm)(4) Engine maximum torque (Nm)(5) SOC(6) Gear command(7) Grade
To reach this decision the first step is to multiply theengine torque limit by three of the gains and compare theresults to the engine load torque These gains are the assistthreshold charge threshold and hysteresis factor Anothergain existing to assist in decision making is the low SOCthreshold which is biased to charging when the SOC dropsbelow it For low SOC mode the more aggressive (chargebias) assistcharge thresholds should be arrived at by takingthe standard values and offsetting them by assistchargethreshold offset Along with comparison of the enginespeed and torque demand with the actual this procedurewill determine the torque demand out of the engine Thisstep determines if the engine is capable of supplying thedemanded power on its own needs assistance from thehybrid or has excess power to be used for charging If theengine is idle and the SOC is less than the idle SOC charge
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
6 Journal of Control Science and Engineering
Pump control pistonSwash plate
Load senseCylinder
Bias spring Pump piston
Pump control valve
Compensator
Implement control valve
Figure 7 Load sensing axial piston pump with swash plate (httpwwwmobilehydraulictipscomefficient-mobile-hydraulic-systems)
Battery
Power electronics
Engine Motorgenerator Transmission Differential
Figure 8 Electric hybrid concept (adopted from [6])
GPS signal
Path identification Cycle identification
Control logicControlsoftware
block
On-machine sensors Hybrid systemactuators
Figure 9 Rule-based control block diagram
that will send out desired torque to the ISG and based onthis the generator-motor actionwill be determined To decidewhether to charge or assist multiple factors are consideredThese factors are represented by gains and thresholds that willenable reaching the decision These gains and thresholds aretuned for different cycles to achieve optimum robust resultsbased on the criteria listed earlierThegains tuned in this logicare delay timer assist threshold charge threshold low SOCthreshold idle charge torque ISG torque-required thresholdengine speed factor assistcharge threshold offset and theidle charge SOC threshold
Table 2 Operator model signals
Cycle model input signals Cycle model output signals(1) Engine speed (rpm) (1) Desired engine speed (rpm)(2) Desired gear number (2) Gear command(3) Lift displacement (m) (3) Lift command(4) Tilt displacement (m) (4) Tilt command(5) Machine ground velocity (5) Brake command
(ms) (6) Bucket load(7) Grade(8) Drawbar
Table 3 Rule-based logic signals
Input signals Output signals(1) Engine speed (rpm) (1) ISG torque command(2) Desired engine speed (rpm)(3) Engine load torque (Nm)(4) Engine maximum torque (Nm)(5) SOC(6) Gear command(7) Grade
To reach this decision the first step is to multiply theengine torque limit by three of the gains and compare theresults to the engine load torque These gains are the assistthreshold charge threshold and hysteresis factor Anothergain existing to assist in decision making is the low SOCthreshold which is biased to charging when the SOC dropsbelow it For low SOC mode the more aggressive (chargebias) assistcharge thresholds should be arrived at by takingthe standard values and offsetting them by assistchargethreshold offset Along with comparison of the enginespeed and torque demand with the actual this procedurewill determine the torque demand out of the engine Thisstep determines if the engine is capable of supplying thedemanded power on its own needs assistance from thehybrid or has excess power to be used for charging If theengine is idle and the SOC is less than the idle SOC charge
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Control Science and Engineering 7
119870-
3000Lock ISG clutch
steer cmd
gear shift modeStep
gear command
engine speed
machine displacement mps
tilt displacement
lift displacement
gear cmnd
drawbar
bucket load
tilt cmd
lift cmd
brake cmd
DesEngSpd
Operator modelgrade engine speed factor
Machine model
In1 Out1Filter2
EngMaxTrq
SOC
EngLoadTrq
Gear Cmd
EngSpd
Des EngSpd In
grade
Tisg
1
119911
ISG and Clutch Logic Simplest
Figure 10 Rulendashbased control upper level Simulink diagram
Mac
hine
velo
city
(ms
)
0 50 100 150 200minus4
minus2
0
2
4
Machine displacement (m)
Simulation
Real machine
Figure 11 Validation machine velocity
Desired
MachineSimulation
0 50 100 150 200500
1000
1500
2000
2500
Machine displacement (m)
Engi
ne sp
eed
(rpm
)
Figure 12 Validation engine speed
threshold then a charge command is issued Charging is alsoenforced when the retard engine speed threshold is exceeded
With any gear upshift or if the engine decelerationrate exceeds the engine deceleration rate threshold assist is
0 50 100 150 2000
02
04
06
08
1
Machine displacement (m)
Con
sum
ed fu
el (l
iters
)
Simulation
Real machine
Figure 13 Validation consumed fuel
triggered The gain tuning process is an iterative and time-consuming process The objective is to obtain an optimalset of parameters for robust design such that slight changesin the gain values do not affect the performance muchThe orthogonal array experiments only require a fractionof the full-factorial combinations and the same result canbe achieved This can be done by using the analysis of thevariance (ANOVA) thus making it the most suitable fortuning the nine gains A predictive model based on theANOM is then formulated through which an estimate ofthe optimal set of interactions is determined and tested Byrepeating this process over time the desired optimal robustsolution can be obtained for each cycle [21]
4 Results
In this section themachine virtual dynamicmodel validationas well as the results obtained during this work is presentedThe validation results show that the used machine model isin good correlation with the real-life machine The results
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
8 Journal of Control Science and Engineering
Baseline machine displacement
Hybrid machine displacement
Baseline machine fuel consumption
Hybrid machine fuel consumption
Hybrid battery state of charge
201510
50
(m)
400300200100
0(gra
ms)
5150494847
()
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
0 5 10 15 20 25 30 35Time (s)
Figure 14 ATL hybrid versus baseline
Baseline machine displacement Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
2010
0
400200
0
5251504948
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
0 5 10 15 20 25 30 35 40Time (s)
(m)
(gra
ms)
()
Figure 15 Moderate truck loading cycle
obtained show that when the engine is downsized the hybridsystem is implemented its control logic gains are optimizedand the system performs very well without GPS or cyclerecognition algorithmThemodel validation is performed byobtaining machine data from the real-life wheel loader andgiving the same operator commands to the machine model
Figure 11 shows both the real and simulated machinevelocitiesThegeneral speed pattern is the same and is in goodagreement The differences could be attributed to the useof approximate dynamics in the different machine systemsFigure 12 shows the desired engine speed sent by the operatorto the machine and implemented in the cycle model thereal machine engine speed and the simulated engine speedFrom Figure 12 it is clear that the modeled engine speedresponse to the desired engine speed command matches to agreat extend the actual engine speed responseThedifferencescould be attributed to the use of approximation and estimatednumbers in various points in themodel including the load onthe machine Figure 13 shows the amount of fuel consumedby the machine versus that estimated by the machine modelThe general trend of the fuel consumption is the same andthe difference in the end point is minimal Thus it canbe concluded that the machine model has a very good
150100
500
1000500
0
5550454035
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
0 10 20 30 40 50 60 70Time (s)
Baseline machine displacement
Hybrid machine displacement
Baseline fuel consumption
Hybrid fuel consumption
Hybrid state of charge
(m)
(gra
ms)
()
Figure 16 Short load and carry cycle
100908070605040302010
0
()
Engine torque contribution
Hybrid torque contribution
0 5 10 15 20 25 30 35Time (s)
Figure 17 Torque contribution
correlation with the real machine Three work cycles wereused in the analysis ATL MTL and SLC cycles Figure 14compares between the baseline regular and the optimizedhybrid wheel loader for the ATL cycle
It is shown that maintaining productivity while decreas-ing the fuel consumption by 20 can be achieved by switch-ing to the downsized hybrid machine Figure 15 comparesbetween the baseline regular and the optimized hybrid wheelloader for the MTL cycle From the obtained results it isshown that maintaining productivity while decreasing thefuel consumption by 265 can be achieved by switching tothe downsized hybrid machine
Figure 16 compares between the baseline regular and theoptimized hybrid wheel loader for the SLC cycle From theobtained results it is shown that increasing productivity by77 and decreasing the fuel consumption by 28 can beachieved by switching to the downsized hybrid machineFrom Figure 17 it can be deduced that the hybrid systemcontributes to about 14ndash17 of the total torque needed by themachine After the optimization of the control parameters foreach independent cycle the investigation of the effect of usingthese parameters in different cycles should be carried outThis investigation will determine if online cycle identificationand GPS mapping of the worksite is needed
The optimized gain of the MTL and SLC will be testedin the ATL and compared to the results obtained by its
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
Journal of Control Science and Engineering 9
016
138
787
053
628
013
773
0
2
4
6
8
10
ATL cycle MTL cycle SLC cycle
Prod
uctiv
ity (
)
ATL gainsMTL gainsSLC gains
minus109 minus14
minus2
Figure 18 Effect of different gain groups on cyclesrsquo productivity
20
239259
22
265 274
208
258282
0
5
10
15
20
25
30
ATL cycle MTL cycle SLC cycle
Fuel
redu
ctio
n (
)
ATL gainsMTL gainsSLC gains
Figure 19 Effect of different gain groups on cyclesrsquo fuel reduction
optimized gains the gain groups of the ATL and SLC will betested in theMTL and compared to the results obtained by itsoptimized gains and the gain groups of theATL andMTLwillbe tested in the SLC and compared to the results obtained byits optimized gains By examining the results closely (Figures18 and 19) all the cycles exhibit a change of less than 2in terms of productivity and less than 3 in terms of fuelreduction from that of the optimized cycle
From this it can be concluded that the knowledge ofthe running cycle is not needed as long as the machineis using any optimized set of parameters Hence onlinecycle identification and GPS mapping is not needed tobe implemented on the machine but remains useful as ananalysis tool for the worksite and operator performance
5 Conclusions
The implementation of an electric hybrid powertrain on amedium wheel loader is expected to maintain the produc-tivity of the machine within acceptable range The usage of
the hybrid system with a downsized engine is expected toreduce the fuel consumption on the machine by 20ndash30 atdifferent cycles The hybrid system supplies 14ndash17 of thetorque required by the machine during its work cycle Byoptimizing the hybrid controller gains for any work cycle thecontroller can be used with any work cycle Thus the GPStracking and gain scheduling algorithms are not needed
Abbreviations
ANOVA Analysis of the varianceATL Aggressive truck-loadingBSFC Brake-specific fuel consumptionCVT Continuously variable transmissionECMS Equivalent consumptions minimization
strategiesGPS Global positioning systemHEV Hybrid electric vehicleISG Integrated starter generatorLLC Long load-and-carryMTL Moderate truck-loadingMWL Medium wheel loaderPHEV Plug-in hybrid electric vehiclePSHEV Power-split hybrid electric vehicleSDP Stochastic dynamic programmingSLC Short load-and-carrySOC State of chargeTC Torque converter
References
[1] C D Rakopoulos and E G GiakoumisDiesel Engine TransientOperation Principles of Operation and Simulation AnalysisSpringer 2009
[2] S CetinkuntMechatronics John Wiley and Sons 2007[3] Z Yafu and C Cheng ldquoStudy on the powertrain for ISG mild
hybrid electric vehiclerdquo in Proceedings of the IEEE Vehicle Powerand Propulsion Conference (VPPC rsquo08) September 2008
[4] G Steinmaurer and L del Re ldquoOptimal energy managementfor mild hybrid operations of vehicle with an integrated startergeneratorrdquo SAE Technical Paper 2005-01-0280 pp 73ndash80 2005
[5] J T B Kessels A J Sijs R M Hermans A A H Damenand P P J van den Bosch ldquoOn-line identification of vehiclefuel consumption for energy and emission management anLTP system analysisrdquo in Proceedings of the American ControlConference (ACC rsquo08) Seattle Wash USA June 2008
[6] J Liu and H Peng ldquoModeling and control of a power-splithybrid vehiclerdquo IEEE Transactions on Control Systems Technol-ogy vol 16 no 6 pp 1242ndash1251 2008
[7] G Paganelli YGuezennec andGRizzoni ldquoOptimizing controlstrategy for hybrid fuel cell vehiclerdquo SAE Technical Paper 2002-01-0102 SAE Warrendale Pa USA 2002
[8] F U Syed M L Kuang M Smith S Okubo and HYing ldquoFuzzy gain-scheduling proportional-integral control forimproving engine power and speed behavior in a hybrid electricvehiclerdquo IEEE Transactions on Vehicular Technology vol 58 no1 pp 69ndash84 2009
[9] M Canova Y Guezennec and S Yurkovich ldquoOn the control ofengine startstop dynamics in a hybrid electric vehiclerdquo Journal
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
10 Journal of Control Science and Engineering
of Dynamic Systems Measurement and Control Transactions ofthe ASME vol 131 no 6 pp 1ndash12 2009
[10] X Lin H Banvait S Anwar and C Yaobin ldquoOptimal energymanagement for a plug-in hybrid electric vehicle real-timecontrollerrdquo in Proceedings of the American Control Conference(ACC rsquo10) pp 5037ndash5042 Baltimore Md USA July 2010
[11] M Gokasan S Bogosyan and D J Goering ldquoSliding modebased powertrain control for efficiency improvement in serieshybrid-electric vehiclesrdquo IEEE Transactions on Power Electron-ics vol 21 no 3 pp 779ndash790 2006
[12] S J Moura H K Fathy D S Callaway and J L Stein ldquoAstochastic optimal control approach for power management inplug-in hybrid electric vehiclesrdquo IEEE Transactions on ControlSystems Technology vol 19 no 3 pp 545ndash555 2011
[13] L Johannesson M Asbogard and B Egardt ldquoAssessing thepotential of predictive control for hybrid vehicle powertrainsusing stochastic dynamic programmingrdquo IEEE Transactions onIntelligent Transportation Systems vol 8 no 1 pp 71ndash83 2007
[14] Y J Huang C L Yin and J W Zhang ldquoDesign of an energymanagement strategy for parallel hybrid electric vehicles usinga logic threshold and instantaneous optimization methodrdquoInternational Journal of Automotive Technology vol 10 no 4pp 513ndash521 2009
[15] S H Lee S D Walters and R J Howlett ldquoIntelligent GPS-based vehicle control for improved fuel consumption andreduced emissionsrdquo in Proceedings of the 12th InternationalConference on Knowledge-Based Intelligent Information andEngineering Systems pp 701ndash708 Zagreb Croatia 2008
[16] Z X Fu ldquoReal-time prediction of torque availability of anipm synchronous machine drive for hybrid electric vehiclesrdquoin Proceedings of the IEEE International Conference on ElectricMachines and Drives pp 199ndash206 San Antonio Tex USA May2005
[17] H A Borhan A Vahidi A M Phillips M L Kuang and I VKolmanovsky ldquoPredictive energy management of a power-splithybrid electric vehiclerdquo in Proceedings of the American ControlConference (ACC rsquo09) pp 3970ndash3976 June 2009
[18] S Grammatico A Balluchi and E Cosoli ldquoA series-parallelhybrid electric powertrain for industrial vehiclesrdquo in Proceed-ings of the IEEEVehicle Power and Propulsion Conference (VPPCrsquo10) pp 1ndash6 September 2010
[19] F Croce Optimal Design of Powertrain and Hydraulic Imple-ment Systems for Construction Equipment Applications [PhDthesis] University of Illinois at Chicago 2010
[20] M Ehsani Y Gao and A Emadi Modern Electric HybridElectric and Fuel Cell Vehicles Fundamentals Theory andDesign Second Edition Taylor amp Francis 2010
[21] Y W Fowlkes and C M Creveling Engineering Methods ForRobust ProductDesignUsing TaguchiMethods in Technology andProduct Development Prentice Hall 1995
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
International Journal of
AerospaceEngineeringHindawi Publishing Corporationhttpwwwhindawicom Volume 2014
RoboticsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Active and Passive Electronic Components
Control Scienceand Engineering
Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
International Journal of
RotatingMachinery
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporation httpwwwhindawicom
Journal ofEngineeringVolume 2014
Submit your manuscripts athttpwwwhindawicom
VLSI Design
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Shock and Vibration
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Civil EngineeringAdvances in
Acoustics and VibrationAdvances in
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Electrical and Computer Engineering
Journal of
Advances inOptoElectronics
Hindawi Publishing Corporation httpwwwhindawicom
Volume 2014
The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014
SensorsJournal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Chemical EngineeringInternational Journal of Antennas and
Propagation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
Navigation and Observation
International Journal of
Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014
DistributedSensor Networks
International Journal of
top related